US12375605B2ActiveUtilityA1

Automated systems for communications analysis according to recording restrictions

65
Assignee: LIVEPERSON INCPriority: Dec 19, 2022Filed: May 31, 2024Granted: Jul 29, 2025
Est. expiryDec 19, 2042(~16.4 yrs left)· nominal 20-yr term from priority
Inventors:Bruce Ramsay
H04M 3/42221G10L 15/22G10L 15/1815G10L 15/063H04M 2203/556H04M 2203/551H04M 3/5175G06Q 30/016G06Q 30/01G06Q 10/20G06Q 10/0639
65
PatentIndex Score
0
Cited by
9
References
21
Claims

Abstract

Disclosed embodiments provide a framework for automatically establishing recording parameters according to specified recording restrictions and generating analytics corresponding to communications recorded subject to the recording restrictions. During a communications session between a user and an agent, a system can identify any recording restrictions corresponding to user communications exchanged during the communications session. The system automatically processes, in real-time, communications exchanged during the communications session as these communications are exchanged to identify the user communications and agent communications. The system generates a transcript that includes the agent communications but selectively records and transcribes the user communications according to the recording restrictions. A machine learning algorithm is trained to generate a set of inferences corresponding to a user sentiment based on historic recordings and transcripts of historic communications sessions between users and agents, as well as corresponding feedback. From the set of inferences, the system generates agent analytics.

Claims

exact text as granted — not AI-modified
What is claimed is: 
     
       1. A computer-implemented method, comprising:
 dynamically processing, in real-time, communications exchanged during a communications session between a user and an agent to isolate user communications associated with the user, wherein the communications session is associated with an intent; 
 continuously generating a recording, wherein the recording includes the user communications and agent communications associated with the agent, and wherein the recording is generated according to an affirmative user consent; 
 detecting a revocation of the affirmative user consent during the communications session, wherein the revocation is communicated through a new user communication exchanged during the communications session; 
 automatically removing the user communications from the recording such that the recording is devoid of the user communications and any subsequent user communications exchanged during the communications session; 
 dynamically training a machine learning algorithm to generate a set of inferences corresponding to user sentiment associated with the intent, wherein the machine learning algorithm is dynamically trained using historic recordings corresponding to historic communications sessions between users and agents, and feedback corresponding to the historic communications sessions; 
 processing the agent communications from the recording through the machine learning algorithm to generate the set of inferences; and 
 generating agent analytics corresponding to the intent, wherein the agent analytics are generated based on the set of inferences and the recording. 
 
     
     
       2. The computer-implemented method of  claim 1 , wherein automatically removing the user communications from the recording further includes:
 inserting one or more audial nulls in the recording to replace the user communications. 
 
     
     
       3. The computer-implemented method of  claim 1 , wherein generating the agent analytics includes:
 identifying a set of agent training needs, wherein the set of agent training needs correspond to actions performable to improve agent responses to intents. 
 
     
     
       4. The computer-implemented method of  claim 1 , wherein detecting the revocation of the affirmative user consent further includes:
 processing the new user communication as the new user communication is exchanged during the communications session; and 
 detecting one or more anchor terms corresponding to a request to cease recording of the user communications. 
 
     
     
       5. The computer-implemented method of  claim 1 , further comprising:
 calculating a voice signature associated with the agent, wherein the voice signature is calculated during an onboarding process; and 
 identifying the user communications based on the user communications not corresponding to the voice signature. 
 
     
     
       6. The computer-implemented method of  claim 1 , further comprising:
 processing the user communications to generate a unique voice signature associated with the user; and 
 using the unique voice signature to automatically identify and remove the user communications from the recording. 
 
     
     
       7. The computer-implemented method of  claim 1 , further comprising:
 activating an agent bot, wherein the agent bot is implemented to solicit the affirmative user consent; and 
 receiving the affirmative user consent, wherein the affirmative user consent is provided in a response to a solicitation generated by the agent bot during the communications session. 
 
     
     
       8. A system, comprising:
 one or more processors; and 
 memory storing thereon instructions that, as a result of being executed by the one or more processors, cause the system to:
 dynamically process, in real-time, communications exchanged during a communications session between a user and an agent to isolate user communications associated with the user, wherein the communications session is associated with an intent; 
 continuously generate a recording, wherein the recording includes the user communications and agent communications associated with the agent, and wherein the recording is generated according to an affirmative user consent; 
 detect a revocation of the affirmative user consent during the communications session, wherein the revocation is communicated through a new user communication exchanged during the communications session; 
 automatically remove the user communications from the recording such that the recording is devoid of the user communications and any subsequent user communications exchanged during the communications session; 
 dynamically train a machine learning algorithm to generate a set of inferences corresponding to user sentiment associated with the intent, wherein the machine learning algorithm is dynamically trained using historic recordings corresponding to historic communications sessions between users and agents, and feedback corresponding to the historic communications sessions; 
 process the agent communications from the recording through the machine learning algorithm to generate the set of inferences; and 
 generate agent analytics corresponding to the intent, wherein the agent analytics are generated based on the set of inferences and the recording. 
 
 
     
     
       9. The system of  claim 8 , wherein the instructions that cause the system to automatically remove the user communications from the recording further cause the system to:
 insert one or more audial nulls in the recording to replace the user communications. 
 
     
     
       10. The system of  claim 8 , wherein the instructions that cause the system to generate the agent analytics further cause the system to:
 identify a set of agent training needs, wherein the set of agent training needs correspond to actions performable to improve agent responses to intents. 
 
     
     
       11. The system of  claim 8 , wherein the instructions that cause the system to detect the revocation of the affirmative user consent further cause the system to:
 process the new user communication as the new user communication is exchanged during the communications session; and 
 detect one or more anchor terms corresponding to a request to cease recording of the user communications. 
 
     
     
       12. The system of  claim 8 , wherein the instructions further cause the system to:
 calculate a voice signature associated with the agent, wherein the voice signature is calculated during an onboarding process; and 
 identify the user communications based on the user communications not corresponding to the voice signature. 
 
     
     
       13. The system of  claim 8 , wherein the instructions further cause the system to:
 process the user communications to generate a unique voice signature associated with the user; and 
 use the unique voice signature to automatically identify and remove the user communications from the recording. 
 
     
     
       14. The system of  claim 8 , wherein the instructions further cause the system to:
 activate an agent bot, wherein the agent bot is implemented to solicit the affirmative user consent; and 
 receive the affirmative user consent, wherein the affirmative user consent is provided in a response to a solicitation generated by the agent bot during the communications session. 
 
     
     
       15. A non-transitory computer-readable storage medium storing thereon executable instructions that, as a result of being executed by one or more processors of a computer system, cause the computer system to:
 dynamically process, in real-time, communications exchanged during a communications session between a user and an agent to isolate user communications associated with the user, wherein the communications session is associated with an intent; 
 continuously generate a recording, wherein the recording includes the user communications and agent communications associated with the agent, and wherein the recording is generated according to an affirmative user consent; 
 detect a revocation of the affirmative user consent during the communications session, wherein the revocation is communicated through a new user communication exchanged during the communications session; 
 automatically remove the user communications from the recording such that the recording is devoid of the user communications and any subsequent user communications exchanged during the communications session; 
 dynamically train a machine learning algorithm to generate a set of inferences corresponding to user sentiment associated with the intent, wherein the machine learning algorithm is dynamically trained using historic recordings corresponding to historic communications sessions between users and agents, and feedback corresponding to the historic communications sessions; 
 process the agent communications from the recording through the machine learning algorithm to generate the set of inferences; and 
 generate agent analytics corresponding to the intent, wherein the agent analytics are generated based on the set of inferences and the recording. 
 
     
     
       16. The non-transitory computer-readable storage medium of  claim 15 , wherein the executable instructions that cause the computer system to automatically remove the user communications from the recording further cause the computer system to:
 insert one or more audial nulls in the recording to replace the user communications. 
 
     
     
       17. The non-transitory computer-readable storage medium of  claim 15 , wherein the executable instructions that cause the computer system to generate the agent analytics further cause the computer system to:
 identify a set of agent training needs, wherein the set of agent training needs correspond to actions performable to improve agent responses to intents. 
 
     
     
       18. The non-transitory computer-readable storage medium of  claim 15 , wherein the executable instructions that cause the computer system to detect the revocation of the affirmative user consent further cause the computer system to:
 process the new user communication as the new user communication is exchanged during the communications session; and 
 detect one or more anchor terms corresponding to a request to cease recording of the user communications. 
 
     
     
       19. The non-transitory computer-readable storage medium of  claim 15 , wherein the executable instructions further cause the computer system to:
 calculate a voice signature associated with the agent, wherein the voice signature is calculated during an onboarding process; and 
 identify the user communications based on the user communications not corresponding to the voice signature. 
 
     
     
       20. The non-transitory computer-readable storage medium of  claim 15 , wherein the executable instructions further cause the computer system to:
 process the user communications to generate a unique voice signature associated with the user; and 
 use the unique voice signature to automatically identify and remove the user communications from the recording. 
 
     
     
       21. The non-transitory computer-readable storage medium of  claim 15 , wherein the executable instructions further cause the computer system to:
 activate an agent bot, wherein the agent bot is implemented to solicit the affirmative user consent; and 
 receive the affirmative user consent, wherein the affirmative user consent is provided in a response to a solicitation generated by the agent bot during the communications session.

Cited by (0)

No later patents cite this yet.

References (0)

No backward citations on record.